For Review Only - University of Toronto T-Space · overview of current estimates across traits and includes correlations with novel traits that constitutes an original addition to

For Review OnlyGenetic Correlations among Selected Traits in Canadian

Holsteins

Journal: Canadian Journal of Animal Science

Manuscript ID CJAS-2018-0190.R2

Manuscript Type: Article

Date Submitted by the Author: 19-Feb-2019

Complete List of Authors: Martin, Pauline; University of Guelph, Centre for Genetic Improvement of Livestock, Animal Biosciences; Institut National de la Recherche Agronomique (INRA), Génétique Animale et Biologie Intégrative (GABI), AgroParisTech, Université Paris-SaclayBaes, Christine; University of Guelph, Centre for Genetic Improvement of Livestock, Animal BiosciencesHoulahan, Kerry; University of Guelph, Centre for Genetic Improvement of Livestock, Animal BiosciencesRichardson, Caeli; University of Guelph, Centre for Genetic Improvement of Livestock, Animal BiosciencesJamrozik, Janusz; University of Guelph, Centre for Genetic Improvement of Livestock, Animal Biosciences; Canadian Dairy NetworkMiglior, Filippo; University of Guelph, Centre for Genetic Improvement of Livestock, Animal Biosciences; Canadian Dairy Network

Keywords: Genetic correlations, Holstein

https://mc.manuscriptcentral.com/cjas-pubs

Canadian Journal of Animal Science

For Review Only

1

Martin et al. – Genetic correlations among traits in Holstein

Genetic Correlations among Selected Traits in Canadian Holsteins

P. Martin1,2*, C. Baes1, K. Houlahan1, C. M. Richardson1, J. Jamrozik1,3, and F. Miglior1,3

1Centre for Genetic Improvement of Livestock, Department of Animal Biosciences, University of

Guelph, Guelph, ON

2Génétique Animale et Biologie Intégrative (GABI), Institut National de la Recherche

Agronomique (INRA), AgroParisTech, Université Paris-Saclay, Jouy en Josas, France

3Canadian Dairy Network, Guelph, ON

1Corresponding author: [email protected]

Abstract

In the Canadian dairy industry, there are currently over 80 traits routinely evaluated, and more are

considered for potential selection. Particularly, in the last few years, recording has commenced for

several new phenotypes required to introduce novel traits with high economic importance into the

selection program. However, without a systematic estimation of the genetic correlations that exist

among traits, the potential results of indirect selection are unknown. Therefore, twenty-nine traits

representative of the trait diversity for first lactation Canadian animals were selected. Their two-

by-two genetic correlations were estimated from a dataset of 62,498 first lactation Holstein cows,

using a Markov Chain Monte Carlo Gibbs sampling approach. The general tendencies among the

Page 1 of 31



mailto:[email protected]

For Review Only

2

groups of traits confirm that production traits are negatively correlated with fertility traits and that

functional traits are positively correlated with one another. The association of udder depth with

fertility and disease resistance has also been highlighted. This contribution offers a comprehensive

overview of current estimates across traits and includes correlations with novel traits that

constitutes an original addition to the literature. These new estimates can be used for newly

developed genomic evaluation models and possibly leads to more accurate estimations of the dairy

cows’ overall genetic merit.

Key Words: Genetic correlations, Holstein

Page 2 of 31



For Review Only

3

In order to be considered for selection in dairy cattle populations, a specific trait must have

an economic value, sufficient genetic variability and heritability, be measurable at a low cost and

be clearly and consistently recordable (Shook 1989). Historically, most selection programs have

largely focused on milk production, with some emphasis on type traits, for their ability to meet

this criteria through their high economic importance and systematic data recording (Miglior et al.

2017). However, strong genetic selection for production traits has resulted in unfavorable, indirect

selection for reduced health and fertility, highlighting the existence of antagonistic genetic

correlations among economically important traits (Miglior et al., 2017). Although large differences

in magnitude exist among studies, there is a general unfavorable relationship reported between

milk production and reproductive performance, (Veerkamp et al. 2001; Kadarmideen et al. 2003;

Pryce et al. 2004; Melendez and Pinedo 2007) as well as between milk production and health traits

(Simianer et al. 1991; Kadarmideen et al. 2000). These results further confirm the importance of

considering genetic correlations when selecting for multiple traits. Weigel et al. (2017) defined

these correlations by how the genetic superiority for one trait tends to be inherited with genetic

superiority or inferiority for another trait. The cause of a genetic correlation may be found at the

genomic level, due to linkage or pleiotropy among the regions influencing the considered traits

(Rauw et al. 1998). Therefore, estimates of genetic correlations are specific to the population under

selection as they are influenced by the allele frequencies of that population (Falconer and Mackay

1996).

Currently, there are over 80 traits routinely evaluated by the Canadian Dairy Network

(Guelph, ON; https://www.cdn.ca/). The introduction of genomic selection, and the development

of additional recording systems and proxies, has permitted the evaluation of new traits that were

previously too challenging to be recorded in the overall population. This contributed to, and even

Page 3 of 31



https://www.cdn.ca/

For Review Only

4

accelerated, the availability of traits under consideration for selection. To improve the selection

process and prepare for potential inclusion of novel traits such as feed efficiency and methane

emission in the composite national indexes, it is imperative to complete a systematic evaluation of

the genetic correlations between current and novel traits. If some genetic correlations have been

previously estimated for Canadian Holstein, most of them remain unknown. The objective of this

paper was to estimate the missing correlations among traits of interest.

MATERIALS AND METHODS

Choice of Traits

An evaluation of the genetic correlations between over 80 traits is computationally

demanding and complex, therefore, a selection was made among the traits. To avoid the

multiplication of traits, the first level of composite trait was used for conformation traits instead

of each individually recorded trait, with the exception of udder depth considering its importance

in the selection objective. Our analysis was limited to first parity cows to take advantage of the

existing literature that mostly focus on primiparous animals. This resulted in removal of disease

traits that have a low occurrence in first lactation, such as milk fever. Finally, a few traits were not

included due to their nature of not being suitable for correlation estimation in our case (longevity

traits, for instance, as we consider only first lactation animals). Overall, 29 of the 80 traits were

selected.

Trait Definitions

Production traits investigated were milk yield (MY), protein yield (PY), fat yield (FY),

protein percent (P%) and fat percent (F%) and were expressed on a 305-day lactation basis. Udder

depth (UD) was defined as a score ranging from 1 to 9, with an intermediate optimum, while other

type traits, mammary system (MS), feet and legs (FL), dairy strength (DS) and rump, were

Page 4 of 31



For Review Only

5

composite traits ranging from 40 to 97. The exact definition of the conformation phenotypes can

be found on the Holstein Canada website

(https://www.holstein.ca/Public/en/Services/Classification/Breakdown_of_Traits). Fertility traits

were split into two categories: those involving fertility prior to first calving, referred to as “heifer”

and those involving fertility during the first lactation, referred to as “cow”. Heifer fertility traits

included age at first service (AFS), 56-day non-return rate (NRR) (0=back in heats, 1=pregnant),

and the interval from first service to conception (FSTC). Cow fertility traits included the interval

from calving to first service (CTFS), NRR, FSTC, and days open (DO). Milking speed (MSP)

and milking temperament (MT) were defined as a score ranging from 1 (very slow / very nervous)

to 5 (very fast / very calm). Calving ease (CE) was defined as a score ranging from 1 (unassisted

or unobserved calving) to 4 (caesarean). Calf survival (CS) was defined as 0 = stillborn within the

first 24h and 1 = alive. The somatic cell score (SCS) was calculated from the test-days occurring

in the first 150 days of lactation. Only animals with at least three different measures were retained

and the cell counts (SCC) of the different test-days were averaged before being log-transformed

to SCS using the formula SCS = log2 (SCC/100 000) +3. Health disorders, clinical mastitis (CM),

displaced abomasum (DA), ketosis, metritis, retained placenta (RP), cystic ovaries (CO) and

lameness were defined as binary traits, where 0 = no case, and 1 = at least one case. An animal

was considered sick if a health event was recorded during the first 305 days after calving.

Population Resources and Phenotypes

Phenotypic data was extracted from the Canadian Dairy Network (Guelph, ON) database.

Holstein cows that calved for the first time from 2000 onwards and had a phenotype for every trait

(with the only possible exception of health traits) were considered. A minimum of 50 cows from

the same herd was required. The health dataset provided by the Canadian Dairy Network contained

Page 5 of 31



https://www.holstein.ca/Public/en/Services/Classification/Breakdown_of_Traits

For Review Only

6

historically recorded health (disease) events; however, event recording was not consistent across

herds or years. Not all the herds started health data recording at the same time and some of them

record only partial information (only mastitis events, for instance). To deal with this heterogeneity

of recording, only herds with at least one health event recorded were selected. Then, to distinguish

between missing information and healthy animals, we considered that the cow was healthy if at

least one health event (other than mastitis) was recorded in the herd during the calving year.

Otherwise, the phenotype was considered missing. The final dataset (after edits) consisted of

62,498 cows from 663 herds and 53,711 cows from the same herds for health data. The number of

cows for the health data set is smaller as recording of health traits started only after 2006.

Pedigree was traced as far back as possible, resulting in a pedigree file with 319,299

animals. Animals with performances came from 5,423 different sires. Animals not related to others

were discarded.

Models and Analyses

Among all considered correlations, 128 of 406 were found in Canadian literature, from 12

different sources (Miglior et al., 2007; Loker et al., 2009; Thomas, 2011; Koeck et al., 2012a; b,

2013a; b, 2015a; b; Jamrozik et al., 2013, 2016; Jamrozik and Kistemaker, 2016). Most of the

correlations were only calculated once, with a few of them found in two different articles. Three

correlations (between CM and somatic cell score (SCS), between DA and ketosis and between

metritis and RP) were estimated in three different articles. As these correlations were previously

estimated, they were not estimated as part of this work. The previously estimated correlations are

reported in Supplementary material.

The correlations were estimated from covariance components using bi-variate linear

animal models. Although threshold models are supposed to be more appropriate for binary traits,

Page 6 of 31



For Review Only

7

it was decided, for a matter of homogenization among variable and with the Canadian literature,

to use also linear models for binary and qualitative traits. Numerous studies have found indeed no

improvement of using threshold models compared to linear models (e.g. Negussie et al. 2008).

The model considered for all traits can be expressed in matrix notation as:

y = Xb + Z1h +Z2a +e

where y is the vector of observations for the trait, b is the vector of fixed effects for the trait, h is

the vector of random effects, a is the vector of animal additive genetic effects, e is the vector of

residuals, and X, Z1 and Z2 are respective incidence matrixes assigning observations to effects.

Random effects were assumed to be normally distributed with means equal to zero and covariance

structure equal to

𝑉𝑎𝑟(ℎ𝑎𝑒) = (𝐼ℎ ⊗ 𝐻 0 0

0 𝐺⨂𝐴 00 0 𝐼𝑟⨂𝑅)

Where G is a (co)variance matrix of random direct additive genetic effects, R is the residual

(co)variance matrix and H is the (co)variance matrix of a potential additional random effect to that

specific trait. The A matrix represents the additive genetic relationships among animals, and Ih and

Ir are identity matrices which have orders equal to levels of appropriate random genetic and

residuals effects.

The environmental effects included in the various models were chosen to emulate those

from the routine national evaluation models considering the specificities of our sampled

population. Fixed and random effects including in each model for the various analyzed traits can

be found in Table 1.

Page 7 of 31



For Review Only

8

Variance components were estimated using a Markov Chain Monte Carlo (MCMC) Gibbs

sampling approach with the RJMC procedure within the DMU package (Madsen and Jensen 2008).

The total length of the Gibbs chain was 2,000,000, with a burn-in of 200,000. Flat prior distribution

was assumed for fixed and random effects and an inverted Wishart distribution was assumed for

variance component estimation. Estimates from a preliminary study performed on a subset of the

dataset were used as priors for variance components. The conservative burn-in period was

determined based on trace plots of selected covariance components, tested with various priors.

This was sufficient to minimize the influence of a potential lack of accuracy from the priors.

Estimates of genetic correlations were calculated as posterior means of all samples after

burn-in. The statistical significance of point estimates was determined using an approximated 95%

Bayesian credible interval, by determining whether 0 was included in the interval or not. The

interval was obtained by excluding the 2.5% more extreme samples of each side from the posterior

distribution. Independent numbers of samples were estimated using the method of initial monotone

sequence estimator (Geyer 1992).

As a validation of correlations estimated in the current study, two correlations that were

already estimated in the literature from Canadian data were re-estimated using the dataset and

methods presented in this paper. These two correlations were the correlation between MY and FY

and the correlation between CTFS and FSTC.

The complete correlation matrix (i.e. including both estimated correlations and correlations

from the Canadian literature) was checked with the R software (R Development Core Team 2005)

and appropriate bending was applied to make it positive definite following Schaeffer’s method

(Schaeffer 2014). Briefly, this method starts from the equation

G=UDU’

Page 8 of 31



For Review Only

9

with G being the correlation matrix and D having the eigenvalues of G on its diagonal. Some

appropriate corrections are performed on the negative eigenvalues to make them positive, based

on the square of the sum of their value and the lowest positive eigenvalue. Then the correlation

matrix is reconstructed using the modified D.

RESULTS AND DISCUSSION

Descriptive Statistics

Descriptive statistics for the analyzed traits, post editing, are presented in Table 2. Restrictions on

herd size and ensuring each animal had a phenotype for all analyzed traits likely introduced a slight

bias. The animals selected for this study had slightly higher production levels than the average

animal. Some differences were observed in the fertility traits, with animals selected for this study

being bred earlier (at a younger age and/or sooner after calving) than the population average. This

difference in fertility traits is not surprising, as the animals selected for this study were from large

herds that recorded phenotypes for all traits. These large herds are often associated with high

adoption of new technologies and specific attention allocated to reproductive performance.

Another point about the data is that only animals phenotyped on all traits were kept in the study.

By doing this, we did not only discard herds with partial phenotyping, but also all animals that

were culled before the end on the observation period because of poor health or poor fertility. For

this reason, we have introduced another bias in the analyses that may have influenced the results

of the correlation estimations.

There was no difference between the selected animals and the population average for

workability or type traits. Frequencies observed for health traits, with the exception of CM, were

lower than those reported by Koeck et al. (2012b), where mean disease frequencies of 12.6 (CM),

3.7 (DA), 4.5 (ketosis), 4.6 (RP), 10.8 (metritis), 8.2 (CO) and 9.2 (lameness), were observed. This

Page 9 of 31



For Review Only

10

suggests that some missing data may have been taken for healthy animals despite our data

validation.

Estimated Correlations

The genetic correlations estimated in this study are presented in Table 3 to Table 6. The

complete matrix, including genetic and residual correlations when estimated, is available in

Supplementary Table 1.

Correlations between production traits and all other traits are presented in Table 3. Previous

Canadian studies estimated the correlations between MY and other production traits

(Supplementary Table 1). These literature estimates were similar to other studies (Dematawewa

and Berger 1998; Mokhtari et al. 2015; Frioni et al. 2017; Gibson and Dechow 2018). Fat yield

and F% were favorably correlated (0.38), while other correlations among FY and PY with F% and

P% were close to 0 (between -0.1 and 0.1). This pattern follows those found in previous studies

(Boichard and Bonaïti 1987; Lembeye et al. 2016).

Genetic correlations between yield and type traits were previously estimated in Canadian

literature (Supplementary Table 1), however, correlations between F% and P% with type traits

were estimated in the current study. These estimates for F% and P% with type traits were less than

0.20, ranging from 0.15 (UD with P%) to -0.03 (FL with P%), the correlations between the milk

contents and UD and MS being the only ones significant. Previous Canadian literature found no

correlation between production yields and FL, and favorable correlations between production

yields with DS and MS. These estimates are in concordance with results between type traits and

production traits recently estimated in US Brown Swiss cattle (Gibson and Dechow 2018).

Page 10 of 31



For Review Only

11

Correlations between yield traits and heifer fertility traits (-0.24 to -0.13 for NRR and 0.01

to 0.28 for FSTC) were weak and unfavorable, i.e. higher yields were correlated to lower NRR

and longer FSTC. Estimates between production yields and cow fertility traits were low with

correlations near zero. However, correlation estimates between production yields and DO, a trait

that accounts for several components of cow fertility, were significantly unfavorable (0.18 to 0.21).

This unfavorable correlation between production and fertility is well-known and has been

previously discussed (Andersen-Ranberg et al. 2005; Abe et al. 2009; Mokhtari et al. 2015; Frioni

et al. 2017; Gibson and Dechow 2018). Some authors have pointed out that the evidence of a strong

antagonist association between milk production and reproductive performance is open to criticism

due to physiological and management factors that are not taken into account in the analyses (Bello

et al. 2012; LeBlanc 2013). Correlations between production traits and CE were unfavorable

(around 0.20 and significant for the yield traits and around 0.05 and not significant for the

contents), in accordance with estimates found by Eaglen et al. (2013).

Correlations between production traits with MSP and MT estimated in this study were near

zero (-0.10 to 0.07), and in the same range as previously found by Gibson and Dechow (2018). A

correlation of 0.39 between CM and MY, found through analysis of Canadian literature, is in the

range reported in reviews by Rupp and Boichard (2003) and Martin et al. (2018). Previous

Canadian literature also provides detail on the correlations between production traits and diseases,

such as DA, RP, CO, ketosis, metritis and lameness. Generally, production traits were favorably

correlated with ketosis and metritis and almost uncorrelated with DA and RP. Milk yield was

unfavorable correlated with CO and lameness.

Table 4 presents correlations of type traits with the remaining traits. Udder depth had a

strong positive correlation with MS (0.80), which was expected considering UD is included in the

Page 11 of 31



For Review Only

12

calculation of MS. Udder depth had a slightly negative correlation with other type traits (-0.17 to

0.09). The correlations between type traits and MS are close to zero (-0.06 to 0.08).

Udder depth was found to be favorably correlated with all fertility traits. Other type traits

such as FL, DS and rump were favorably correlated with heifer fertility, while unfavorably

corelated with cow fertility. Correlations between fertility and MS were, however, close to zero.

These results follow previous work done on the associations between type traits and fertility traits

(Zink et al. 2011; Gibson and Dechow 2018).

Milking speed was found favorably correlated with both MS (0.17) and UD (0.14), in line

with current literature (Wiggans et al. 2007; Gibson and Dechow 2018). A favorable correlation (-

0.15) was also found between UD and resistance to various disease (DA, ketosis and CO), which

could indicate shallow udders are genetically correlated with lower incidence of disease. Feet and

legs score was favorably correlated with lameness (-0.46), and DS was unfavorably associated

with displaced abomasum (0.26). The unfavorable correlation between DS and displaced

abomasum was also found by Dechow et al., (2004). All the correlations reported in this paragraph

were significantly different from 0.

Correlations between reproduction traits, workability traits and health traits are presented

in Table 5. Days open ranged in correlation with other reproduction traits from -0.20 for NRR in

cow to 0.88 for FSTC in cow. However, only the correlations of DO with its components (CTFS

and FSTC), which were both above 0.8, were significant. Other studies have estimated correlations

between various measures of fertility (e.g. Veerkamp et al., 2001; Andersen-Ranberg et al., 2005;

Abe et al., 2009; Mokhtari et al., 2015). Although the phenotypes considered in these studies are

diverse and no direct comparison can be made; the general trends are consistent with those found

in this study. Unfavorable correlations were observed between CM and the three heifer fertility

Page 12 of 31



For Review Only

13

traits: AFS (-0.04), NRR (0.20) and FSTC (-0.41), only the last one being significantly different

from 0. The direction is the same in the Canadian literature for the correlations between these

fertility traits and SCS. This could indicate that heifers who become pregnant at a younger age

may be genetically predisposed to having mastitis. The opposite was observed for correlations

between CM and SCS with cow fertility, where the correlations were favorable (0.29, -0.19 and

0.20 for the correlation between CM and CTFS, NRR and FSTC respectively) but not significant

due to their large credible intervals. This is one of the rare cases where we seems to have such a

difference between heifer and cow fertility traits in their correlations with other traits. Even though

they were not always significant, correlations of heifer and cow fertility traits with production

traits or UD and MS showed the same trend, for example. Nevertheless, mastitis often occurs

around calving time or in early lactation, the same periods when cow fertility is challenged. Heifers

and cows are not facing the same biological needs at the reproduction time, as cows are just

recovering from calving while heifers are not. It is therefore not surprising that correlations with

CM and SCS are in the opposite direction between cows and heifers. The difference of trend

observed for the correlation with RP tends to confirm this hypothesis, even though these

correlations are not significant.

Estimates of correlations between workability traits, calving traits, and health traits are

presented in Table 6. Workability traits were strongly positively correlated (0.58, significant). This

indicates that cows with a genetically faster MSP also seem to have a calmer MT. However, this

tendency was also slightly correlated with a higher risk of CM and SCS (significant correlations

around 0.2). The relationship between MSP with SCS and CM has been previously analyzed in

various studies and confirms the association found here (Boettcher et al., 1998; Rupp and

Boichard, 1999; Zwald et al., 2005 for SCS; Govignon-Gion et al., 2012; Pérez-Cabal and

Page 13 of 31



For Review Only

14

Charfeddine, 2013 for CM). Some studies, however, found evidence that animals with low MSP

may be more susceptible to CM (Lund et al. 1994; Rupp and Boichard 2003). Moreover, Samoré

and Groen (2010) found a nonlinear relationship between EBV for SCS and MSP. This suggests

that the relationship between MSP and CM traits may not be simply linear and requires more

investigation. Slight associations were found between workability traits and other disease traits,

all less than 0.20 in magnitude. Genetically, fast milking animals appear to be significantly more

resistant to RP (-0.17) and CO (-0.15). Calm milking temperament appears slightly but

significantly genetically correlated with less lameness (-0.07).

Some significant correlations were also observed between calving traits and health traits,

such as difficult calving being associated with DA (0.18), or stillborn calves being associated with

RP (-0.62). It is known from the literature that cases of RP occur around calving, and that almost

all cases of DA occur during the first 100 days of lactation (Zwald et al. 2004; Koeck et al. 2012a).

In early lactation, cows have large physiological demands while going through various transitional

changes (Sordillo et al. 2009). At this time, cows are in a negative energy balance and metabolic

diseases may follow as a consequence of a severe and prolonged period of time in this state

(Collard et al. 2000).

Accuracy of the Estimates

Among all the Gibbs sampling analyses, the number of independent samples for genetic

correlations ranged from 182 to 7,213, with an average of 738 and a standard deviation of 1103.

The number of independent samples varied among traits, with the highest numbers being found

for production traits and the lowest for fertility traits. These numbers were in the same range as

some previously mentioned in the literature (Steinbock et al. 2003; Jamrozik et al. 2005).

Page 14 of 31



For Review Only

15

As we previously mentioned, two correlations that were already estimated in the literature

from Canadian data were re-estimated in this study for validation purposes. A correlation of 0.76

with a 95% Bayesian credible interval of [0.75; 0.77] between milk yield and fat yield and a

correlation of 0.47 [0.21; 0.71] between calving to first service and first service to conception in

cow, were determined. These values are a slightly higher than Canadian estimates from the

literature, 0.57 and 0.31, respectively. If the difference between the literature and the correlations

estimated in the current study for calving to first service and first service to conception is not

significant, the correlation estimated by Miglior et al. (2007) for milk yield and fat yield does not

fall into our posterior interval. This slight gap may be explained by differences between the two

datasets. Analysis in Miglior et al. (2007) was performed on test day records from Holstein cows

only in the province of Quebec. The dataset for this study was a sample of Holstein cows across

Canada that considered a 305-day lactation. Both values are in the range of what was found in the

literature (Boichard and Bonaïti 1987; Dematawewa and Berger 1998; Mokhtari et al. 2015;

Lembeye et al. 2016; Gibson and Dechow 2018).

The size of the interval was highly variable, depending on the traits considered. It ranged

from 0.04 among production traits to 1.42 between NRR (heifer) and lameness. These results were

in the same ranges and sometimes slightly higher than what was found in literature (Jamrozik et

al. 2005; Abe et al. 2009; Zink et al. 2011; Koeck et al. 2012a, 2013b; Mokhtari et al. 2015;

Jamrozik et al. 2016). The magnitude of the credible intervals for some correlations made these

results difficult to interpret. Several factors can influence the accuracy of genetic correlation

estimation. Accuracy is dependent on the dataset including sample size, quality of phenotyping,

and repeated measures. The models used, and the nature of the traits themselves also have

influence on the accuracy of the prediction. For these reasons, Falconer and Mackay (1996)

Page 15 of 31



For Review Only

16

observed that estimates of genetic correlations are usually subject to large sampling errors. In the

current study, the less accurate estimations were found between fertility traits and health traits.

Fertility traits have a very low heritability (Jamrozik and Kistemaker 2016) and health traits,

despite careful editing criteria, probably contain recording errors. Moreover, incidences of disease

were low when compared to the number of animals in the population. This may explain why

accuracy was low for those traits. Even though the estimates have large intervals, these correlations

have never been estimated before from Canadian data and represent important information.

Considering that we used a 95% credible interval to determine the possible range of the

correlation real value, there is, by definition, a 5% chance that the true correlation is outside this

interval. As no less than 278 correlations were estimated in the current study, it could be expected

that around 14 correlations are outside the credible interval that was predicted.

Considered in its entirety –i.e. including the estimates from the Canadian literature

(Supplementary Table 1), the correlation matrix was of rank 29 and was not positive definite, as

six eigenvalues were negative. Correlation matrices calculated from a single sample are supposed

to be positive definite. In our case, the matrix is a patchwork of values coming from different

sources and it is therefore not surprising that negative eigenvalues are present. Following the

method of Schaeffer (2014), the matrix was transformed to become positive definite. This

corrected matrix is presented in Supplementary Table 2. Overall, the uncorrected and the corrected

matrices were very similar, with a Pearson correlation of 0.99 between them, and the average

absolute difference being 0.02 ± 0.03.

Page 16 of 31



For Review Only

17

CONCLUSIONS

This study focused on estimating correlations among selected traits for the Canadian

Holstein cattle population. Correlations that were not previously reported in Canadian literature

over the past ten years between the selected traits were estimated. This study provides correlations

that have not previously been estimated and will be used specifically in applications of selection

index studies, breeding strategies, and estimated response of selection. Future work using multiple

trait models with more than two traits at a time may provide more accurate estimates to be used

for selection programs.

ACKNOWLEDGMENTS

We gratefully acknowledge funding by the Efficient Dairy Genome Project, funded by Genome

Canada (Ottawa, Canada), Genome Alberta (Calgary, Canada), Ontario Genomics (Toronto,

Canada), Alberta Ministry of Agriculture (Edmonton, Canada), Ontario Ministry of Research and

Innovation (Toronto, Canada), Ontario Ministry of Agriculture, Food and Rural Affairs (Guelph,

Canada), Canadian Dairy Network (Guelph, Canada), GrowSafe Systems (Airdrie, Canada),

Alberta Milk (Edmonton, Canada), Victoria Agriculture (Australia), Scotland's Rural College

(Edinburgh, UK), USDA Agricultural Research Service (United States), Qualitas AG

(Switzerland), Aarhus University (Denmark). Funding from Alberta Innovates Technology

Futures and the Ontario Centres of Excellence (Ontario Network of Entrepreneurs ONE) is also

acknowledged.

Page 17 of 31



For Review Only

18

Page 18 of 31



For Review Only

19

REFERENCES

Abe, H., Masuda, Y., and Suzuki, M. 2009. Relationships between reproductive traits of heifers and cows and yield traits for Holsteins in Japan. J. Dairy Sci. 92: 4055–4062. doi:10.3168/jds.2008-1896.

Andersen-Ranberg, I.M., Klemetsdal, G., Heringstad, B., and Steine, T. 2005. Heritabilities, Genetic Correlations, and Genetic Change for Female Fertility and Protein Yield in Norwegian Dairy Cattle. J. Dairy Sci. 88: 348–355. doi:10.3168/jds.S0022-0302(05)72694-1.

Bello, N.M., Stevenson, J.S., and Tempelman, R.J. 2012. Invited review: Milk production and reproductive performance: Modern interdisciplinary insights into an enduring axiom. J. Dairy Sci. 95: 5461–5475. doi:10.3168/jds.2012-5564.

Boettcher, P.J., Dekkers, J.C.M., and Kolstad, B.W. 1998. Development of an Udder Health Index for Sire Selection Based on Somatic Cell Score, Udder Conformation, and Milking Speed. J. Dairy Sci. 81: 1157–1168. doi:10.3168/jds.S0022-0302(98)75678-4.

Boichard, D., and Bonaïti, B. 1987. Genetic parameters for first lactation dairy traits in Friesian, Montbéliarde and Normande breeds. Génétique Sélection Évolution 19: 337–350. doi:10.1051/gse:19870306.

Chang, Y.M., Andersen-Ranberg, I.M., Heringstad, B., Gianola, D., and Klemetsdal, G. 2006. Bivariate analysis of number of services to conception and days open in Norwegian red using a censored threshold-linear model. J. Dairy Sci. 89: 772–778. doi:10.3168/jds.S0022-0302(06)72138-5.

Collard, B.L., Boettcher, P.J., Dekkers, J.C.M., Petitclerc, D., and Schaeffer, L.R. 2000. Relationships Between Energy Balance and Health Traits of Dairy Cattle in Early Lactation. J. Dairy Sci. 83: 2683–2690. doi:10.3168/jds.S0022-0302(00)75162-9.

Dechow, C.D., Rogers, G.W., Sander-Nielsen, U., Klei, L., Lawlor, T.J., Clay, J.S., Freeman, A.E., Abdel-Azim, G., Kuck, A., and Schnell, S. 2004. Correlations Among Body Condition Scores from Various Sources, Dairy Form, and Cow Health from the United States and Denmark. J. Dairy Sci. 87: 3526–3533. doi:10.3168/jds.S0022-0302(04)73489-X.

Dematawewa, C.M., and Berger, P.J. 1998. Genetic and phenotypic parameters for 305-day yield, fertility, and survival in Holsteins. J. Dairy Sci. 81: 2700–2709. doi:10.3168/jds.S0022-0302(98)75827-8.

Eaglen, S.A.E., Coffey, M.P., Woolliams, J.A., and Wall, E. 2013. Direct and maternal genetic relationships between calving ease, gestation length, milk production, fertility, type, and lifespan of Holstein-Friesian primiparous cows. J. Dairy Sci. 96: 4015–4025. doi:10.3168/jds.2012-6229.

Falconer, D.S., and Mackay, T.F.C. 1996. Introduction to Quantitative Genetics. 4th ed. Longman, UK.

Frioni, N., Rovere, G., Aguilar, I., and Urioste, J.I. 2017. Genetic parameters and correlations between days open and production traits across lactations in pasture based dairy production systems. Livest. Sci. 204: 104–109. doi:10.1016/j.livsci.2017.08.018.

Geyer, C.J. 1992. Practical Markov Chain Monte Carlo. Stat. Sci. 7: 473–483. doi:10.1214/ss/1177011137.

Gibson, K.D., and Dechow, C.D. 2018. Genetic parameters for yield, fitness, and type traits in US Brown Swiss dairy cattle. J. Dairy Sci. 101: 1251–1257. doi:10.3168/jds.2017-13041.

Page 19 of 31



For Review Only

20

Govignon-Gion, A., Dassonnevill, R., Baloche, G., and Ducrocq, V. 2012. Genetic evaluation of mastitis in dairy cattle in France. Interbull Bull. 46: 121–126.

Jamrozik, J., Fatehi, J., Kistemaker, G.J., and Schaeffer, L.R. 2005. Estimates of Genetic Parameters for Canadian Holstein Female Reproduction Traits. J. Dairy Sci. 88: 2199–2208. doi:10.3168/jds.S0022-0302(05)72895-2.

Jamrozik, J., and Kistemaker, G.J. 2016. Updated Genetic Parameters for Holstein Reproductive Traits Using More Recent Data. [Online] Available: https://www.cdn.ca/Articles/GEBAPR2016/3_New%20RP%20Parameters%20-%20Janusz.pdf [2018 Sep. 26].

Jamrozik, J., Koeck, A., Kistemaker, G.J., and Miglior, F. 2016. Multiple-trait estimates of genetic parameters for metabolic disease traits, fertility disorders, and their predictors in Canadian Holsteins. J. Dairy Sci. 99: 1990–1998. doi:10.3168/jds.2015-10505.

Jamrozik, J., Koeck, A., Miglior, F., Kistemaker, G.J., Schenkel, F.S., Kelton, D.F., and Van Doormaal, B.J. 2013. Genetic and Genomic Evaluation of Mastitis Resistance in Canada. Interbull Bull. 47: 9pp.

Kadarmideen, H.N., Thompson, R., Coffey, M.P., and Kossaibati, M.A. 2003. Genetic parameters and evaluations from single- and multiple-trait analysis of dairy cow fertility and milk production. Livest. Prod. Sci. 81: 183–195. doi:10.1016/S0301-6226(02)00274-9.

Kadarmideen, H.N., Thompson, R., and Simm, G. 2000. Linear and threshold model genetic parameters for disease, fertility and milk production in dairy cattle. Anim. Sci. 71: 411–419. doi:10.1017/S1357729800055338.

Koeck, A., Jamrozik, J., Kistemaker, G.J., Schenkel, F.S., Kelton, D.F., and Miglior, F. 2015a. Estimation of genetic parameters for fertility disorders and their predictors in Canadian Holsteins. [Online] Available: https://www.cdn.ca/Articles/GEBAPR2015/7B_DCBGC%20March%202015%20-%20Astrid%20Koeck%20-%20Fertility%20Disorders%20Evaluation.pdf [2018 Sep. 26].

Koeck, A., Jamrozik, J., Kistemaker, G.J., Schenkel, F.S., Moore, R.K., Lefebvre, D.M., Kelton, D.F., and Miglior, F. 2015b. Estimation of genetic parameters for metabolic traits and their predictors in Canadian Holsteins. [Online] Available: https://www.cdn.ca/Articles/GEBAPR2015/7_DCBGC%20March%202015%20-%20Astrid%20Koeck%20-%20Metabolic%20Diseases%20Evaluation.pdf [2018 Sep. 26].

Koeck, A., Loker, S., Miglior, F., Kelton, D.F., Jamrozik, J., and Schenkel, F.S. 2014. Genetic relationships of clinical mastitis, cystic ovaries and lameness with milk yield and somatic cell score in first-lactation Canadian Holsteins. J. Dairy Sci. 97: 5806–5813. doi:10.3168/jds.2013-7785.

Koeck, A., Miglior, F., Chapinal, N., Kelton, D.F., and Schenkel, F.S. 2013a. Genetic associations between lameness and feet and leg conformation traits in Holsteins. Res Rep DCBGC Febr. 2013: 8pp.

Koeck, A., Miglior, F., Jamrozik, J., Kelton, D.F., and Schenkel, F.S. 2013b. Genetic associations of ketosis and displaced abomasum with milk production traits in early first lactation of Canadian Holsteins. J. Dairy Sci. 96: 4688–4696. doi:10.3168/jds.2012-6408.

Koeck, A., Miglior, F., Kelton, D.F., and Schenkel, F.S. 2012a. Health recording in Canadian Holsteins: data and genetic parameters. J. Dairy Sci. 95: 4099–4108. doi:10.3168/jds.2011-5127.

Page 20 of 31



For Review Only

21

Koeck, A., Miglior, F., Kelton, D.F., and Schenkel, F.S. 2012b. Short communication: Genetic parameters for mastitis and its predictors in Canadian Holsteins. J. Dairy Sci. 95: 7363–7366. doi:10.3168/jds.2012-5648.

LeBlanc, S.J. 2013. Is a high level of milk production compatible with good reproductive performance in dairy cows? Anim. Front. 3: 84–91. doi:10.2527/af.2013-0038.

Lembeye, F., Lopez-Villalobos, N., Burke, J.L., and Davis, S.R. 2016. Estimation of genetic parameters for milk traits in cows milked once- or twice-daily in New Zealand. Livest. Sci. 185: 142–147. doi:10.1016/j.livsci.2016.01.022.

Liu, A., Lund, M.S., Wang, Y., Guo, G., Dong, G., Madsen, P., and Su, G. 2017. Variance components and correlations of female fertility traits in Chinese Holstein population. J. Anim. Sci. Biotechnol. 8: 56. doi:10.1186/s40104-017-0189-x.

Loker, S., Bastin, C., Miglior, F., Sewalem, A., Schaeffer, L.R., Jamrozik, J., Ali, A., and Osborn, V. 2012. Genetic and environmental relationships between body condition score and milk production traits in Canadian Holsteins. J. Dairy Sci. 95: 410–419. doi:10.3168/jds.2011-4497.

Lund, T., Miglior, F., Dekkers, J.C.M., and Burnside, E.B. 1994. Genetic relationships between clinical mastitis, somatic cell count, and udder conformation in Danish Holsteins. Livest. Prod. Sci. 39: 243–251. doi:10.1016/0301-6226(94)90203-8.

Madsen, P., and Jensen, J. 2008. An User’s Guide to DMU. A package for analyzing multivariate mixed models. Version 6, release 4.7. Danish Institute of Agricultural Sciences, Tjele, Denmark.

Martin, P., Barkema, H.W., Brito, L.F., Narayana, S.G., and Miglior, F. 2018. Symposium review: Novel strategies to genetically improve mastitis resistance in dairy cattle. J. Dairy Sci. 101: 2724–2736. doi:10.3168/jds.2017-13554.

Melendez, P., and Pinedo, P. 2007. The association between reproductive performance and milk yield in Chilean Holstein cattle. J. Dairy Sci. 90: 184–192. doi:10.3168/jds.S0022-0302(07)72619-X.

Miglior, F., Fleming, A., Malchiodi, F., Brito, L.F., Martin, P., and Baes, C.F. 2017. A 100-Year Review: Identification and genetic selection of economically important traits in dairy cattle. J. Dairy Sci. 100: 10251–10271. doi:10.3168/jds.2017-12968.

Miglior, F., Sewalem, A., Jamrozik, J., Bohmanova, J., Lefebvre, D.M., and Moore, R.K. 2007. Genetic analysis of milk urea nitrogen and lactose and their relationships with other production traits in Canadian Holstein cattle. J. Dairy Sci. 90: 2468–2479. doi:10.3168/jds.2006-487.

Mokhtari, M.S., Moradi Shahrbabak, M., Nejati Javaremi, A., and Rosa, G.J.M. 2015. Genetic relationship between heifers and cows fertility and milk yield traits in first-parity Iranian Holstein dairy cows. Livest. Sci. 182: 76–82. doi:10.1016/j.livsci.2015.10.026.

Negussie, E., Strandén, I., and Mäntysaari, E.A. 2008. Genetic analysis of liability to clinical mastitis, with somatic cell score and production traits using bivariate threshold–linear and linear–linear models. Livest. Sci. 117: 52–59. doi:10.1016/j.livsci.2007.11.009.

Pérez-Cabal, M.A., and Charfeddine, N. 2013. Genetic relationship between clinical mastitis and several traits of interest in Spanish Holstein dairy cattle. Interbull Bull. 47: 77–81.

Pryce, J.E., Royal, M.D., Garnsworthy, P.C., and Mao, I.L. 2004. Fertility in the high-producing dairy cow. Livest. Prod. Sci. 86: 125–135. doi:10.1016/S0301-6226(03)00145-3.

Page 21 of 31



For Review Only

22

R Development Core Team 2005. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. [Online] Available: http://www.R-project.org.

Rauw, W.M., Kanis, E., Noordhuizen-Stassen, E.N., and Grommers, F.J. 1998. Undesirable side effects of selection for high production efficiency in farm animals: a review. Livest. Prod. Sci. 56: 15–33. doi:10.1016/S0301-6226(98)00147-X.

Rupp, R., and Boichard, D. 1999. Genetic parameters for clinical mastitis, somatic cell score, production, udder type traits, and milking ease in first lactation Holsteins. J. Dairy Sci. 82: 2198–2204. doi:10.3168/jds.S0022-0302(99)75465-2.

Rupp, R., and Boichard, D. 2003. Genetics of resistance to mastitis in dairy cattle. Vet. Res. 34: 671–688. doi:10.1051/vetres:2003020.

Samoré, A.B., and Groen, A.F. 2010. Proposal of an udder health genetic index for the Italian Holstein Friesian based on first lactation data. Ital. J. Anim. Sci. [Online] Available: http://agris.fao.org/agris-search/search.do?recordID=DJ2012044424 [2018 Mar. 15].

Schaeffer, L.R. 2014. Making covariance matrices positive definite. [Online] Available: http://www.aps.uoguelph.ca/~lrs/ELARES/PDforce.pdf [2018 Apr. 23].

Shook, G.E. 1989. Selection for Disease Resistance. J. Dairy Sci. 72: 1349–1362. doi:10.3168/jds.S0022-0302(89)79242-0.

Simianer, H., Solbu, H., and Schaeffer, L.R. 1991. Estimated Genetic Correlations Between Disease and Yield Traits in Dairy Cattle. J. Dairy Sci. 74: 4358–4365. doi:10.3168/jds.S0022-0302(91)78632-3.

Sordillo, L.M., Contreras, G.A., and Aitken, S.L. 2009. Metabolic factors affecting the inflammatory response of periparturient dairy cows. Anim. Health Res. Rev. 10: 53–63. doi:10.1017/S1466252309990016.

Steinbock, L., Näsholm, A., Berglund, B., Johansson, K., and Philipsson, J. 2003. Genetic Effects on Stillbirth and Calving Difficulty in Swedish Holsteins at First and Second Calving. J. Dairy Sci. 86: 2228–2235. doi:10.3168/jds.S0022-0302(03)73813-2.

Thomas, A.D. 2011. Study of health traits and relative economic values using simulation. MSc thesis, University of Guelph, Guelph, Canada.

Veerkamp, R.F., Koenen, E.P.C., and Jong, G.D. 2001. Genetic Correlations Among Body Condition Score, Yield, and Fertility in First-Parity Cows Estimated by Random Regression Models. J. Dairy Sci. 84: 2327–2335. doi:10.3168/jds.S0022-0302(01)74681-4.

Weigel, K.A., VanRaden, P.M., Norman, H.D., and Grosu, H. 2017. A 100-Year Review: Methods and impact of genetic selection in dairy cattle—From daughter–dam comparisons to deep learning algorithms. J. Dairy Sci. 100: 10234–10250. doi:10.3168/jds.2017-12954.

Wiggans, G.R., Thornton, L.L.M., Neitzel, R.R., and Gengler, N. 2007. Short Communication: Genetic Evaluation of Milking Speed for Brown Swiss Dairy Cattle in the United States. J. Dairy Sci. 90: 1021–1023. doi:10.3168/jds.S0022-0302(07)71587-4.

Zink, V., Štípková, M., and Lassen, J. 2011. Genetic parameters for female fertility, locomotion, body condition score, and linear type traits in Czech Holstein cattle. J. Dairy Sci. 94: 5176–5182. doi:10.3168/jds.2010-3644.

Zwald, N.R., Weigel, K.A., Chang, Y.M., Welper, R.D., and Clay, J.S. 2004. Genetic Selection for Health Traits Using Producer-Recorded Data. I. Incidence Rates, Heritability Estimates, and Sire Breeding Values. J. Dairy Sci. 87: 4287–4294. doi:10.3168/jds.S0022-0302(04)73573-0.

Page 22 of 31



For Review Only

23

Zwald, N.R., Weigel, K.A., Chang, Y.M., Welper, R.D., and Clay, J.S. 2005. Genetic Evaluation of Dairy Sires for Milking Duration Using Electronically Recorded Milking Times of Their Daughters. J. Dairy Sci. 88: 1192–1198. doi:10.3168/jds.S0022-0302(05)72785-5.

Page 23 of 31



For Review Only

1

Table 1 Fixed and random effects fitted in the genetic parameter estimation for various trait groups

Fixed Effects Random Effect

Herd -year -season

Age and region at calving

DIM2 at test day Sex of calf

Herd-classification

round-classifier

Stage of lactation-age

at calving

Region - birth year - birth month

Age at prev. calving-

month of 1st insemination

Age at prev. calving–month of

prev. calving

Herd-year of birth

Production Traits (MY, FY, PY, F%, P%) + + - - - - - - - -

Type Traits (UD, MS, FL, DS, Rump) - - - - + + - - - -

Heifer Fertility Traits (AFS, FSTCh, NRRh) - - - - - - + - - +

CTFS - - - - - - + - ++

Other Cow Fertility Traits (NRRc, FSTCc,

DO)- - - - - - + + - +

Workability Traits (MSP, MT) + + + - - - - - - -

Calving Traits (CE, CS) + + - + - - - - - -

Health Traits (CM, SCS, DA, Ketosis, Metritis, RP, CO, Lameness)

+ + - - - - - - - -

Note: A + indicated the inclusion of the effect in the model and a – indicates the effect was not included in the model.DIM: days in milk, MY: milk yield, FY: fat yield, PY: protein yield, F%: fat percent, P%: protein percent, UD: udder depth, MS: mammary system, FL: feet and legs, DS: dairy strength, AFS: age at first service, NNR: 56-day non-return rate, FSTC: first service to conception, CTFS: calving to first service, DO: days open, MSP: milking speed, MT: milking temperament, CE: calving ease, CS: calf survival, CM: clinical mastitis, SCS: somatic cell score, DA: displaced abomasum, RP: retained placenta, CO: cystic ovaries

Page 24 of 31



For Review Only

1

Page 25 of 31



For Review Only

2

Table 2. Descriptive statistics of analyzed data for first lactation Holstein cows

Trait Number of records Mean Standard Deviation Minimum MaximumMilk (kg) 62,498 9,000.17 1,455.10 5,677 12,629Fat (kg) 62,498 341.34 56.57 213 482

Protein (kg) 62,498 288.06 43.15 187 392Fat Percent (%) 62,498 3.82 0.44 2.81 4.90

Protein Percent (%) 62,498 3.21 0.19 2.79 3.69Udder Depth (points) 62,498 4.95 1.38 1 9

Mammary System (score) 62,498 79.45 4.99 40 89Feet and Legs (score) 62,498 79.16 5.04 40 89Dairy Strength (score) 62,498 81.15 3.83 40 93

Rump (score) 62,498 81.02 5.33 40 91Age at First Service (days) 62,498 465.16 48.02 365 606

Non-Return Rate Heifer (0/1) 62,498 0.71 0.46 0 1First Service to Conception Heifer

(days) 62,498 18.16 32.57 0 145

Calving to First Service (days) 62,498 76.21 19.66 43 142Non-Return Rate Cow (0/1) 62,498 0.57 0.49 0 1

First Service to Conception Cow (days) 62,498 25.09 33.35 0 127Days Open (days) 62,498 102.12 37.58 48 214

Milking Speed (points) 62,498 3.09 0.73 1 5Milking Temperament (points) 62,498 3.26 0.78 1 5

Calving Ease (points) 62,498 1.56 0.69 1 4Calf Survival (0/1) 62,498 0.89 0.32 0 1

Clinical Mastitis (0/1) 53,711 0.13 0.34 0 1Somatic Cell Score (score) 62,498 2.01 1.28 0.11 5.96Displaced Abomasum (0/1) 53,711 0.03 0.16 0 1

Ketosis (0/1) 53,711 0.02 0.15 0 1Metritis (0/1) 53,711 0.05 0.22 0 1

Retained Placenta (0/1) 53,711 0.04 0.19 0 1Cystic Ovaries (0/1) 53,711 0.05 0.22 0 1

Lameness (0/1) 53,711 0.07 0.26 0 1

Page 26 of 31



For Review Only

1

Page 27 of 31



For Review Only

2

Table 3. Genetic correlations for the estimated production traits with all the other traits (with the 95% Bayesian credible interval in square brackets and correlations significantly different from 0 in bold)

MY FY PY F% P%

F% - 0.38[0.36 ; 0.40]

-0.10[-0.12 ; -0.07] - -

P% - 0.08[0.06 ; 0.10]

0.07[0.05 ; 0.09] - -

UD - - - 0.09[0.05 ; 0.12]

0.15[0.12 ; 0.19]

MS - - - 0.08[0.05 ; 0.13]

0.07[0.02 ; 0.11]

FL - - - 0.04[-0.04 ; 0.13]

0.04[-0.04 ; 0.12]

DS - - - 0.03[-0.01 ; 0.08]

-0.03[-0.07 ; 0.01]

Rump - - - -0.01[-0.05 ; 0.04]

0.00[-0.04 ; 0.05]

AFS (heifer) - - - 0.01[-0.11 ; 0.10]

-0.12[-0.22 ; -0.03]

NRR (heifer) -0.20[-0.41 ; 0.00]

-0.13[-0.34 ; 0.09]

-0.24[-0.45 ; -0.04]

0.12[-0.08 ; 0.34]

0.09[-0.09 ; 0.27]

FSTC (heifer) 0.23[0.01 ; 0.47]

0.01[-0.24 ; 0.26]

0.28[0.05 ; 0.51]

-0.30[-0.51 ; -0.10]

-0.07[-0.27 ; 0.14]

CTFS (cow) - - - -0.02[-0.13 ; 0.10]

-0.15[-0.26 ; -0.04]

NRR (cow) -0.07[-0.26 ; 0.12]

-0.02[-0.22 ; 0.17]

-0.21[-0.40 ; -0.04]

0.06[-0.12 ; 0.22]

-0.20[-0.37 ; -0.05]

FSTC (cow) 0.05[-0.15 ; 0.25]

-0.04[-0.25 ; 0.16]

0.11[-0.10 ; 0.29]

-0.06[-0.24 ; 0.12]

0.13[-0.05 ; 0.30]

DO (cow) 0.18[0.06 ; 0.30]

0.18[0.05 ; 0.30]

0.21[0.09 ; 0.33]

-0.01[-0.12 ; 0.10]

0.03[-0.08 ; 0.13]

MSP 0.06[0.02 ; 0.10]

0.07[0.04 ; 0.11]

0.06[0.02 ; 0.09]

0.03[-0.01 ; 0.06]

-0.01[-0.04 ; 0.02]

MT - - - -0.05[-0.09 ; -0.01]

-0.04[-0.07 ; 0.00]

CE 0.19[0.12 ; 0.26]

0.22[0.15 ; 0.29]

0.21[0.14 ; 0.28]

0.06[0.00 ; 0.14]

0.04[-0.03 ; 0.10]

CS -0.03[-0.14 ; 0.07]

-0.11[-0.21 ; -0.01]

-0.08[-0.18 ; 0.02]

-0.11[-0.20 ; -0.01]

-0.09[-0.18 ; 0.00]

CM - -0.11[-0.20 ; -0.02]

-0.03[-0.12 ; 0.06]

-0.12[-0.21 ; -0.03]

-0.03[-0.11 ; 0.06]

DA - - - -0.02[-0.11 ; 0.07]

-0.06[-0.15 ; 0.02]

Ketosis - - - -0.01[-0.09 ; 0.07]

-0.17[-0.25 ; -0.10]

Metritis - - - -0.03[-0.07 ; 0.01]

-0.04[-0.08 ; 0.00]

RP - - - 0.02[-0.08 ; 0.13]

0.00[-0.09 ; 0.10]

CO - - - -0.02[-0.13 ; 0.08]

-0.16[-0.26 ; -0.07]

Lameness - - - -0.01[-0.05 ; 0.03]

-0.07[-0.10 ; -0.03]

Note: MY: milk yield, FY: fat yield, PY: protein yield, F%: fat percent, P%: protein percent, UD: udder depth, MS: mammary system, FL: feet and legs, DS: dairy strength, AFS: age at first service, NNR: 56-day non-return rate, FSTC: first service to conception, CTFS: calving to first service, DO: days open, MSP: milking speed, MT: milking temperament, CE: calving ease, CS: calf survival, CM: clinical mastitis, DA: displaced abomasum, RP: retained placenta, CO: cystic ovaries

Page 28 of 31



For Review Only

Table 4. Genetic correlations for the estimated type traits with all remaining traits (with the 95% Bayesian credible interval in square brackets and correlations significantly different from 0 in bold)

UD MS FL DS RumpMS 0.80

[0.76 ; 0.83]FL -0.17

[-0.27 ; -0.07]-0.06

[-0.19 ; 0.07]DS -0.10

[-0.16 ; -0.04]0.08

[0.00 ; 0.15]0.32

[0.21 ; 0.43]Rump -0.09

[-0.16 ; -0.03]0.00

[-0.08 ; 0.08]0.05

[-0.07 ; 0.18]0.32

[0.24 ; 0.38]AFS (heifer) -0.01

[-0.12 ; 0.11]0.06

['-0.08 ; 0.21]-0.02

[-0.20 ; 0.17]-0.26

[-0.39 ; -0.13]-0.14

[-0.28 ; 0.00]NRR

(heifer)0.15

[-0.07 ; 0.38]0.05

[-0.20 ; 0.31]0.12

[-0.16 ; 0.42]0.23

[0.00 ; 0.48]0.35

[0.13 ; 0.57]FSTC

(heifer)-0.17

[-0.41 ; 0.09]-0.09

[-0.36 ; 0.21]-0.24

[-0.52 ; 0.10]-0.15

[-0.40 ; 0.12]-0.26

[-0.50 ; 0.03]CTFS (cow) -0.22

[-0.33 ; -0.10]-0.07

[-0.22 ; 0.08]-0.11

[-0.29 ; 0.08]0.35

[0.21 ; 0.49]0.14

[-0.01 ; 0.29]NRR (cow)

0.19[0.01 ; 0.40]

0.15[-0.06 ; 0.39]

-0.36[-0.58 ; -0.12]

-0.11[-0.31 ; 0.10]

0.02[-0.19 ; 0.25]

FSTC (cow) -0.25[-0.45 ; -0.04]

-0.14[-0.36 ; 0.11]

0.30[0.04 ; 0.55]

0.40[0.18 ; 0.64]

0.10[-0.11 ; 0.35]

DO (cow) -0.13[-0.25 ; -0.01]

-0.02[-0.16 ; 0.14]

0.14[-0.04 ; 0.35]

0.47[0.33 ; 0.62]

0.17[0.02 ; 0.32]

MSP 0.14[0.09 ; 0.20]

0.17[0.08 ; 0.21]

0.07[-0.05 ; 0.20]

-0.06[-0.13 ; 0.01]

-0.09[-0.17 ; -0.02]

MT 0.01[-0.05 ; 0.07]

0.03[-0.06 ; 0.11]

0.02[-0.12 ; 0.15]

0.00[-0.08 ; 0.07]

0.03[-0.05 ; 0.11]

CE -0.12[-0.23 ; -0.03]

-0.05[-0.18 ; 0.07]

0.09[-0.10 ; 0.25]

0.05[-0.07 ; 0.16]

0.13[0.00 ; 0.24]

CS -0.10[-0.23 ; 0.01]

-0.06[-0.21 ; 0.08]

0.05[-0.15 ; 0.22]

0.14[0.00 ; 0.27]

0.11[-0.04 ; 0.25]

CM - -0.12[-0.27 ; 0.02]

0.05[-0.18 ; 0.27]

0.10[-0.04 ; 0.24]

0.06[-0.08 ; 0.21]

SCS - -0.29[-0.38 ; -0.20]

0.03[-0.12 ; 0.17]

0.05[-0.04 ; 0.13]

0.01[-0.09 ; 0.10]

DA -0.15[-0.27 ; -0.05]

-0.18[-0.31 ; -0.04]

0.05[-0.13 ; 0.23]

0.26[0.13 ; 0.38]

0.12[-0.02 ; 0.25]

Ketosis -0.15[-0.26 ;-0.04]

-0.16[-0.31 ; -0.02]

-0.14[-0.35 ; 0.08]

0.01[-0.12 ; 0.14]

0.06[-0.09 ; 0.20]

Metritis 0.05[-0.01 ; 0.12]

0.02[-0.07 ; 0.11]

-0.08[-0.23 ; 0.07]

-0.05[-0.13 ; 0.03]

-0.07[-0.16 ; 0.02]

RP 0.14[0.00 ; 0.26]

0.13[-0.03 ; 0.29]

-0.07[-0.27 ; 0.14]

-0.06[-0.20 ; 0.09]

0.00[-0.16 ; 0.15]

CO -0.15[-0.28 ; -0.02]

-0.13[-0.28 ; 0.03]

-0.04[-0.25 ; 0.18]

-0.02[-0.17 ; 0.122]

0.13[-0.03 ; 0.28]

Lameness 0.08[0.01 ; 0.14]

0.00[-0.10 ; 0.09] - -0.02

[-0.10 ; 0.07]-0.02

[-0.11 ; 0.07]Note: UD: udder depth, MS: mammary system, FL: feet and legs, DS: dairy strength, AFS: age at first service, NNR: 56-day non-return rate, FSTC: first service to conception, CTFS: calving to first service, DO: days open, MSP: milking speed, MT: milking temperament, CE: calving ease, CS: calf survival, CM: clinical mastitis, SCS: somatic cell score, DA: displaced abomasum, RP: retained placenta, CO: cystic ovaries

Page 29 of 31



For Review Only

1

Table 5. Genetic correlations for the estimated fertility traits with all the remaining traits (with the 95% Bayesian credible interval in square brackets and correlations significantly different from 0 in bold)

AFS NNR (heifer)

FSTC(heifer)

CTFS NRR(cow)

FSTC(cow)

DO

DO 0.17[-0.03 ; 0.33]

0.14[-0.13 ; 0.36]

0.10[-0.18 ; 0.34]

0.84[0.74 ; 0.94]

-0.20[-0.42 ; 0.05]

0.88[0.79 ; 0.95] -

MSP 0.00[-0.16 ; 0.15]

0.15[-0.09 ; 0.36]

0.19[-0.10 ; 0.46]

0.05[-0.10 ; 0.22]

-0.10[-0.31 ; 0.16]

0.02[-0.25 ; 0.28]

0.09[-0.07 ; 0.26]

MT 0.04[-0.13 ; 0.21]

0.21[-0.05 ; 0.44]

-0.10[-0.36 ; 0.16]

0.18[0.00 ; 0.38]

0.16[-0.07 ; 0.43]

-0.15[-0.42 ; 0.11]

0.05[-0.13 ; 0.24]

CE - - - - - - 0.00[-0.24 ; 0.25]

CS - - - - - - -0.07[-0.31 ; 0.19]

CM -0.04[-0.34 ; 0.25]

0.20[-0.17 ; 0.58]

-0.41[-0.72 ; -0.02]

0.29[0.00 ; 0.64]

-0.19[-0.57 ; 0.24]

0.20[-0.19 ; 0.58]

0.18[-0.13 ; 0.54]

SCS -0.24[-0.40 ; -0.08]

0.04[-0.23 ; 0.28]

-0.07[-0.35 ; 0.23]

0.17[-0.01 ; 0.35]

0.08[-0.17 ; 0.31]

0.39[0.10 ; 0.68]

0.23[0.05 ; 0.40]

DA - 0.07[-0.19 ; 0.33]

0.11[-0.16 ; 0.39] - -0.28

[-0.52 ; -0.01]0.26

[0.01 ; 0.053]0.21

[0.00 ; 0.43]Ketosis - 0.28

[-0.39 ; 0.78]0.13

[-0.59 ; 0.78] - -0.53[-0.88 ; 0.08]

-0.33[-0.84 ; 0.32]

0.72[-0.02 ; 0.96]

Metritis - - 0.51[-0.10 ; 0.93] - - 0.36

[-0.31 ; 0.388]0.53

[-0.15 ; 0.90]RP - - -0.25

[-0.54 ; 0.13] - - 0.18[-0.13 ; 0.49]

0.14[-0.12 ; 0.42]

CO - - 0.03[-0.35 ; 0.43] - - 0.31

[-0.11 ; 0.77]0.34

[0.04 ; 0.70]Lameness - -0.01

[-0.55 ; 0.51]0.35

[-0.33 ; 0.90] - 0.27[-0.54 ; 0.88]

0.41[-0.30 ; 0.90]

0.80[0.15 ; 0.98]

Note: AFS: age at first service, NNR: 56-day non-return rate, FSTC: first service to conception, CTFS: calving to first service, DO: days open, MSP: milking speed, MT: milking temperament, CM: clinical mastitis, SCS: somatic cell score, CE: calving ease, CS: calf survival, DA: displaced abomasum, RP: retained placenta, CO: cystic ovaries

Page 30 of 31



For Review Only

Table 6. Genetic correlations for the estimated workability traits, calving traits and mammary health traits with the remaining traits (with the 95% Bayesian credible interval in square brackets and correlations significantly different from 0 in bold)

MSP MT CE CS CM SCS

MT 0.58[0.52 ; 0.63] - - - - -

CE -0.10[-0.20 ; 0.00]

0.11[-0.01 ; 0.22] - - - -

CS -0.01[-0.14 ; 0.12]

0.08[-0.07 ; 0.22] - - - -

CM 0.20[0.06 ; 0.34]

0.27[0.13 ; 0.41]

-0.01[-0.20 ; 0.19]

0.04[-0.16 ; 0.24] - -

SCS 0.22[0.14 ; 0.29]

0.11[0.03 ; 0.19]

0.07[-0.19 ; 0.06]

0.03[-0.12 ; 0.17] - -

DA 0.08[-0.05 ; 0.20]

0.04[-0.09 ;0.16]

0.18[0.01 ; 0.35]

0.14[-0.05 ; 0.32] - 0.01

[-0.13 ; 0.15]

Ketosis -0.04[-0.16 ; 0.08]

-0.01[-0.14 ; 0.11]

-0.12[-0.32 ; 0.09]

-0.10[-0.22 ; 0.21] - -0.04

[-0.18 ; 0.10]

Metritis 0.08[0.01 ; 0.15]

0.08[0.01 ; 0.15]

0.09[-0.03 ; 0.22]

-0.07[-0.24 ; 0.10] - 0.04

[-0.06 ; 0.13]

RP -0.17[-0.31 ; -0.03]

-0.09[-0.23 ; 0.06]

-0.01[-0.21 ; 0.21]

-0.62[-0.77 ; -0.44] - -0.06

[-0.22 ; 0.10]

CO -0.15[-0.30 ; -0.01]

0.04[-0.11 ; 0.19]

0.01[-0.19 ; 0.23]

0.09[-0.12 ; 0.31] - -

Lameness 0.04[-0.03 ; 0.10]

-0.07[-0.14 ; -0.01]

0.01[-0.11 ; 0.13]

-0.09[-0.24 ; 0.07] - -

Note: MSP: milking speed, MT: milking temperament, CM: clinical mastitis, SCS: somatic cell score, CE: calving ease, CS: calf survival, DA: displaced abomasum, RP: retained placenta, CO: cystic ovaries

Page 31 of 31



Documents

For Review Only - University of Toronto T-Space · overview of current estimates across traits and includes correlations with novel traits that constitutes an original addition to